Read errors on a 2.5 years old HDD: do I need to replace it or is it some diagnostic error?

[n0stalghia]

I've got a 2.5 years old 8TB Western Digital that has had three errors today when reading from a sector. This has never happened before.

 

I'm not experienced in this. Is this some small random error? Is the drive dying? It's not been used too much, so this comes as a surprise to me. Are there some tests that I should run on it? Is there a shell script I can execute to try and read from those specific sectors again to confirm that the error is persistent?

 

The disk in question is disk1 in the diagnostics.

uxie-diagnostics-20210712-1628.zip

Edited July 12, 2021 by n0stalghia

[trurl]

The usually monitored SMART attributes look OK, but syslog seems to indicate an actual disk problem.

 

Run an extended SMART test on the disk.

[n0stalghia]

4 minutes ago, trurl said:

The usually monitored SMART attributes look OK, but syslog seems to indicate an actual disk problem.

 

Run an extended SMART test on the disk.

Thanks, the fast test reported no errors, the extended one will be done in 14 hours.

Edited July 12, 2021 by n0stalghia

[n0stalghia]

48 minutes ago, trurl said:

The usually monitored SMART attributes look OK, but syslog seems to indicate an actual disk problem.

 

Run an extended SMART test on the disk.

SMART test is still running, EDIT: Just realized that this message also killed the extended SMART test. Errors occured - Check SMART report.

 

I just got this message about "current pending sector" being 8. I assume nothing good? The drive is 2.17 years old, so should still be under warranty (3 years).

 

EDIT: Added a second SMART report, this one only contains one file for some reason.

uxie-smart-20210712-1725.zip

Edited July 12, 2021 by n0stalghia

[n0stalghia]

The drive's serial number is WDC_WD80EFAX-68KNBN0_VAGY3XYL - is this one of the 7200 RPM drives that were marketed as 5400 RPM? My parity has a different number, seems to be unaffected.

Edited July 12, 2021 by n0stalghia

[trurl]

14 minutes ago, n0stalghia said:

the extended SMART test. Errors occured

replace

[n0stalghia]

1 minute ago, trurl said:

replace

Can the drive handle one more parity check, or should I power down and replace immediately? I'd lose like 3 TV series, so no big deal (all important data is backed up to the cloud every Monday)

[JonathanM]

1 hour ago, n0stalghia said:

Can the drive handle one more parity check, or should I power down and replace immediately? I'd lose like 3 TV series, so no big deal (all important data is backed up to the cloud every Monday)

That statement does not compute. Parity is realtime, a check just verifies that all drives can be read accurately and the realtime parity has stayed in sync.

 

If everything is working properly you should be able to replace the drive at any time and not lose anything.

[n0stalghia]

10 minutes ago, jonathanm said:

That statement does not compute. Parity is realtime, a check just verifies that all drives can be read accurately and the realtime parity has stayed in sync.

 

If everything is working properly you should be able to replace the drive at any time and not lose anything.

I was not aware that parity is real-time, I always assumed it computes it during check. TIL. Well this is great news.

 

On a "fun" side note, I tried to get a list of all data using find . -print > files.txt and the file got corrupted and threw even more mistakes - I took the server offline and already got an RMA procedure running.

 

Thank you for the help.